Ok, one more pg_autovacuum patch (Thanks to Tom for fixing the long long overflow problem).

This patch addresses two minor issues (the last two pg_autovacuum bugs I know of):

First the CountAtLastAnalyze varable was initially being set incorrectly causing pg_autovacuum to perform unneeded analyzes on during startup.

Second, there was a debug print statement left in that was printing relisshared.

The attached patch fixes both issues, and should be applied to both 7.4 and HEAD.

Thanks,

Matthew O'Connor

*** pg_autovacuum.c.orig	2003-12-07 15:25:08.000000000 -0500
--- pg_autovacuum.c	2003-12-07 15:46:07.000000000 -0500
***************
*** 108,114 ****
  
  	new_tbl->CountAtLastAnalyze =
  		(atol(PQgetvalue(res, row, PQfnumber(res, "n_tup_ins"))) +
! 		 atol(PQgetvalue(res, row, PQfnumber(res, "n_tup_upd"))));
  	new_tbl->curr_analyze_count = new_tbl->CountAtLastAnalyze;
  
  	new_tbl->CountAtLastVacuum =
--- 108,115 ----
  
  	new_tbl->CountAtLastAnalyze =
  		(atol(PQgetvalue(res, row, PQfnumber(res, "n_tup_ins"))) +
! 		 atol(PQgetvalue(res, row, PQfnumber(res, "n_tup_upd"))) +
! 		 atol(PQgetvalue(res, row, PQfnumber(res, "n_tup_del"))));
  	new_tbl->curr_analyze_count = new_tbl->CountAtLastAnalyze;
  
  	new_tbl->CountAtLastVacuum =
***************
*** 120,126 ****
  	new_tbl->reltuples = atoi(PQgetvalue(res, row, PQfnumber(res, "reltuples")));
  	new_tbl->relpages = atoi(PQgetvalue(res, row, PQfnumber(res, "relpages")));
  
- 	log_entry(PQgetvalue(res, row, PQfnumber(res, "relisshared")));
  	if (strcmp("t", PQgetvalue(res, row, PQfnumber(res, "relisshared"))))
  		new_tbl->relisshared = 0;
  	else
--- 121,126 ----

---------------------------(end of broadcast)---------------------------
TIP 4: Don't 'kill -9' the postmaster

Reply via email to